ZIP codes vs. census tracts vs. property-level risk

Why ZIP codes mislead insurance underwriting, how census tracts and boundary-free risk maps differ, and a worked example from Santa Clarita, California.

Updated November 2026

ZIP CODECENSUS TRACTSNO BOUNDARIESZIP 91350: Very HighAbout 40% of residents: moderatePockets of opportunity
In short

ZIP codes were designed to deliver mail, not to measure risk. Rating or restricting insurance by ZIP code assigns one value to an area inside which risk can range from low to extreme. Census tracts are smaller and more uniform; a property-level view, with no boundaries at all, shows the risk where it actually is.

Three ways to draw the same place

ZIP codeCensus tractProperty level
Made forRouting mailPublishing census statisticsDescribing one location
Who defines itUS Postal ServiceUS Census BureauThe data itself; no boundary
What one value coversThousands of people, often very different terrainA neighborhood-sized areaOne address or parcel
What it hidesMost of the variation in riskVariation within the tractNothing at the resolution of the data
Typical use in insuranceRating territories, blanket moratoriumsRegional reporting, finer restrictionsRisk selection, polygon moratoriums

A worked example: ZIP 91350, Santa Clarita

ZIP code 91350 in Santa Clarita, California has a population of 9 to 10 thousand and is under moratorium at several insurers. Measured as one unit, its median wildfire risk is Very High, and adding historically burned areas to the map appears to confirm the decision.

Redrawn by census tract, about 40% of the population turns out to live in moderate-risk areas, which is an acceptable level of risk: only 10% of Moderate risk areas have burned in the last 80 years, against 48% of Very High risk areas. Removing boundaries altogether shows further pockets of lower risk. The opportunity lost to the ZIP-wide view is as high as 50%.

The maps for each step are in Using ZIP codes costs you, and here is why.

What to use instead

  • For reporting, break the portfolio down to census tract rather than stopping at postal code.
  • For concentration, look at exposure street by street. See what exposure management is.
  • For restrictions, draw the area that is actually above your threshold as a polygon. See underwriting moratoriums.
  • For individual decisions, score the address itself. See risk selection.

It is not only ZIP codes

The same problem applies to every administrative boundary used as a proxy for risk: counties, cities, states. They are convenient because data is published for them, and they were useful when the technology to do better did not exist. Hail shows the same pattern at national scale: La Niña raises hail frequency overall, yet hundreds of ZIP codes see fewer events in La Niña years, as shown in Why should insurers look at La Niña differently?

Common questions

Is a ZIP code a geographic area?

Not strictly. A ZIP code is a set of mail delivery routes defined by the US Postal Service. The areas usually drawn on maps are approximations of those routes.

What is a census tract?

A small statistical subdivision of a county defined by the US Census Bureau and designed to be relatively stable over time. Tracts are much smaller and more uniform than most ZIP codes, which makes them a better unit for comparing risk.

Why do insurers still use ZIP codes?

Habit, and convenience: historical data, rating plans and regulatory filings are often organized by ZIP code. The cost is accepting or blocking properties because of their neighbors rather than their own risk.

Next step

See it on your own portfolio.

Book a 30-minute walkthrough tailored to your portfolio and perils of interest.

Request a Demo →